Given the symptoms where two nodes correctly failover but lose communication with the third node after an SFP+ is removed, the first action should be to:
* Validate the SFP+ is fully plugged into the node and switch: This is critical because a loose or partially connected SFP+ can cause intermittent network issues, which seems to be the problem based on the scenario described.
References:
* Nutanix Bible (https://www.nutanixbible.com), particularly the network troubleshooting sections.
* Nutanix Field Installation Guide, which provides detailed steps on how to ensure hardware components like SFPs are correctly installed.

When encountering timeout issues during Foundation using a customer-provided network switch, one common cause can be mismatched or unsupported network configurations, particularly concerning MTU sizes:
* A. Check jumbo frames support of the network switch: Ensuring that the switch supports jumbo frames (MTU of 9000) and is properly configured for it can prevent timeout issues that might arise due to packet fragmentation or dropped packets.
References:
* Nutanix Bible (https://www.nutanixbible.com), which discusses network prerequisites for successful Foundation.
* Nutanix University: NCS-Core 6.8 Training (https://university.nutanix.com), covering the technical setup and network requirements for cluster deployments.

For licensing reasons and to ensure compliance with software license terms, particularly for databases like SQL and Oracle RAC, VM-Host affinity policies are used. These policies help in pinning VMs to specific hosts. In this scenario, SQL Servers must be confined to specific hosts to manage licensing costs effectively, and similarly, Oracle RAC servers must be on separate hosts. The VM-Host affinity policy allows for such configuration by defining rules that keep VMs on specific hosts or separate them as needed.References:
* Nutanix Bible
* Nutanix University NCS-Core Materials